SAP Commits to Apache Iceberg and Expands Databricks Partnership

What happened

SAP has committed to Apache Iceberg, the open table format used to manage large analytical datasets, and has expanded its partnership with Databricks. According to The Register, the moves point to a coming contest over the "context layer", the part of the enterprise data stack that gives raw data its business meaning for analytics and AI.

The headline frames the announcement as a promise to open enterprise data flows in both directions. The reporting indicates that the commitment is about interoperability: data held within SAP's ecosystem should be easier to use in other platforms, and data from elsewhere should be easier to bring into SAP.

Why it matters

For organisations running large data estates, the choice of table format and platform partner determines how easily information can move between systems. An open format such as Iceberg lowers the cost of moving data between engines, which weakens the lock-in that has historically shaped enterprise software purchasing. Expanded partnerships also signal how vendors expect to compete as AI workloads depend more heavily on governed, well-described data.

For leaders in digital transformation, the practical question is less about which vendor wins the context layer and more about how their own teams and customers will find, trust and act on data. Interoperability changes how quickly new AI tools can be put to use, and how much of an organisation's operating model depends on a single supplier.

The Renascence take

Most coverage of data openness will focus on connectors and technical compatibility. The more important question is who owns the meaning attached to the data, because the party that defines context shapes every downstream experience, from a service agent's screen to an AI assistant's answer. Open formats make data portable, but they do not make it understandable.

Treat open table formats as a service-design decision, not just an architecture one. Map the moments where a customer or employee needs a trusted answer, then trace which systems supply the context behind it. Where that context sits with a single vendor, the switching cost is a behavioural trap that will be felt by users long before it appears on a balance sheet. Build the context layer around the experience you want to deliver, and keep the option to change its engine.